W36 - Hitting the Essence

Last week I read a book — "Hitting the Essence." It can be summarized in two parts. First it explains "essence," meaning the fundamental attributes of things, the root of problems, and the underlying logic behind phenomena. Then it addresses "hitting" the essence, offering a method summarized as boldly hypothesize and cautiously verify.

The book is full of methodologies and models. Fortunately, having practiced fundamentals for two years, I could map many abstractions to concrete situations. Reading it felt rewarding; I was often jolted awake. Many fuzzy, habitual ways of thinking and methods I use daily found clear expressions and time‑tested models and methodologies in the book.

For example, I realized that the natural way I troubleshoot production issues can largely be summarized by the "Muller five methods."

Another example is how to define things: what makes a good definition. In daily life I relied more on feeling plus logic, but I couldn't clearly state what that feeling or logic actually was. The book cites Aristotle’s definition of definition: definition = genus (shared) + differentia (distinct). Reviewing several examples made things suddenly clear. For instance, one can define a human as a featherless, bipedal animal, or as a rational animal.

I also think I found the prototype of Amazon’s growth flywheel — the system loop diagram. The book explains finding root causes through systemic analysis. Its theoretical basis is systems theory, and many concepts align. The primary analytical tool used is the system loop diagram, a diagram for showing how a system actually operates. From a system loop diagram you can read what elements make up a system, whether relationships among elements form reinforcing loops or balancing loops, what the input/output pendulum is, and what the delays are.

In form it closely resembles Amazon’s growth flywheel; after careful comparison, the growth flywheel appears to be a subset of the system loop diagram. The growth flywheel focuses only on reinforcing loops: input metrics represent positive input swings while balancing loops are omitted. For example, certain actions improve user experience and thereby drive market share growth. Often this is a chain of reinforcing logic, but market share growth is also affected by external factors like market size — a balancing logic that is hidden.

I understand the reason for doing this is to make the tool more usable by focusing on the core, concentrating only on the core logic of product growth. If you draw the complete system, its increased complexity sharply raises the barrier to understanding. The tool itself would be limited in expression and communication: if a seed lacks low‑cost ways to spread, no matter how robust it is, it cannot sprout everywhere.